Methodology for the analysis of student behavior and performance in an online course

Although many researchers have studied student performance prediction in online courses, they have primarily focused on courses with a linear structure, where students complete lessons and assessments sequentially. However, non-linear courses allow students to take lessons and assessments in any order, making performance prediction more challenging due to varying cumulative assessment percentages among students at any given time. This master's thesis aims to develop a data-driven method for early student performance prediction in non-linear courses. We created a feature extractor and evaluated three types of features: engagement, behavior, and performance. The data comes from Moodle courses designed to prepare high school students for a public university entrance exam. Our method achieved early predictions at 20% of cumulative weight assessment with an F1-score of 0.73 for binary classification and an R² of 0.40 for regression. We also conducted a feature importance analysis, showing that performance and behavior features are the most significant predictors, with engagement features, such as time spent on educational resources, also contributing significantly. In addition to predicting student performance, we performed a clustering analysis and identified four patterns that consistently appear across various cumulative weight assessments. These patterns significantly impact performance and can help educators provide better feedback and more personalized attention to students' needs.

• Autores: Mercado Agudelo, Jhon Fredy
• Fecha de publicación: 2025
• Institución: Universidad de Antioquia
• Repositorio: Repositorio UdeA
• Idioma: eng
• OAI Identifier: oai:bibliotecadigital.udea.edu.co:10495/46166
• Acceso en línea: https://hdl.handle.net/10495/46166
• Palabra clave: Machine learning; Aprendizaje automático; Data mining; MOOCs (Web-based instruction); Cursos en línea masivos en abierto; Student Performance Prediction; http://id.loc.gov/authorities/subjects/sh85079324; http://id.loc.gov/authorities/subjects/sh97002073; http://id.loc.gov/authorities/subjects/sh2013002540
• Rights: openAccess
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/